Bug fix for parallel mode.
Description:
H5FD_mpio_tas_allsame was called for all cases, even when
MPIO is not used for access. That corrupted the internal
file handle structure.
Solution:
Define a macro, IS_H5FD_MPIO(f), for testing if f is opened with
MPIO access. Will call H5FD_mpio_tas_allsame only if this
condition is true.
Platform tested:
O2K, both -64 and -n32 modes.
RELEASE file was brought up to date.
This file is under construction for 1.3.x (1.4.0?) release.
All major changes MUST be reflected in one of the following sections of this file
* New Features
* Bug Fixes
* Platfoms Tested
* Known Problems
Please update New Features section when you have time!!!!!!!!!!!!!
HISTORY file was brought up to date. It contains release information for
HDF5 1.2.0 and 1.2.1 along with the historical changes to Alpha and Beta
versions of HDF5 1.0.X.
Bug Fix.
Description:
Tests failed if drivers multi or split is used (split is implemented
as a special case of multi). When a file created via those two drivers
is opened again, the file sizes retrieved are incorrect. These seemed
to happen to non-little-endian machines only (sun, sgi failed, but
linux succeed.)
Solution:
The error was because the superblock encoding code was keeping track
in nseen the pairs of HADDR needed to be converted but did not tell the
H5Tconvert 2*nseen to convert. Thus only the first half of the addresses
were converted, the rest were left alone. Corrected the numbers of
addresses to convert.
Platforms tested:
IRIX64, Solaris 2.7, Linux.
macros in the dumper. this was basically adding a space after most of the '}'.
there are some places that don't have the space because those '}' were not printed
by the end_obj macro
the header information in the dumper. the header is the part before the data
block.
in h5dump.c there is support for the new -xml option. prints the output in xml
using the structure we just defined in h5tools.h.
h5dump.h- changed the begin_obj and end_obj macros. take an extra parameter now
Bug fix.
Description
When converting a floating point zero to a different format (e.g. Cray),
the mantissa part was not set to zeros complete. Cray does not
seem to mind since the exponent part was set to all 0's. Somehow,
the HDF4 does not like it and convert that pattern to a non-zero
value in IEEE.
Solution
corrected the code to set the whole mantissa to zeros completely.
Platform tested:
J90 & Solaris 2.7.
Groups.html
Added note that object names (in Groups.html) or dataset
names (in Datasets.html) must be unique.
Changed 'boot block' to 'super block'. (Groups.html only)
Files.html
Minor formatting fixes.
Datatypes.html
H5Tconvert: Added plist_id parameter.
References.html
Deleted "not implemented" and related messages.
Added description of H5Rget_object_type.
Deleted all material about internal dataset region references.
[Unlikely to be implemented (QK) and there was very little info.]
Commented out references to H5R_BADTYPE. [This appeared in the
list of "valid types", which it is not. And user is instructed
to test for 'a negative value' rather than 'H5R_BADTYPE'.
So this was deemed not to be user-level information.]
Changed 'boot block' to 'super block'.
Other minor editing changes.
H5Dopen: Corrected description of 'loc_id' parameter.
RM_H5T.html
H5Tset_strpad:
Corrected string padding values list (replaced 'H5T_STR_NULL &
H5T_STR_SPACE' with ''H5T_STR_NULLTERM, 'H5T_STR_NULLPAD, &
H5T_STR_SPACEPAD'.
Added notes re conversion from shorter to longer strings and
from longer to shorter strings.
Other minor edits and formatting.
H5Tget_class
Corrected H5T_ENUM and H5T_REFERENCE entries in class
identifiers table (they were reversed).
Tools.html
h5dump: Added note that '-bb' option is not yet implemented.
Changed 'boot block' to 'super block'.
Corrected link to HDF5 FTP archives.
H5.intro.html
Added paragraph stating that changes in bugfix Releases 1.2.x
are listed in the RELEASE, HISTORY, and README files.
